matt grime

Your idea is on the right lines, though you need to specify more than just p_n goes to x^n: where does 1/p_n go? what about 4, where is that sent?